Using Web Analytics in the Newsroom

Document Sample
Using Web Analytics in the Newsroom Powered By Docstoc
					Measuring Web Success in the Newsroom

            Web Analytics
                for
              Editors


        Version 1 – July 2008



                 by
             Dana Chinn
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________



Contents

Using Web Analytics                                                         2

Developing a Web Analytics Plan                                             3

Defining Newsroom Key Performance Indicators                                4

Basic Behavioral Metrics                                                    5

Attitudinal Research Areas                                                  6

Key Performance Indicators

      Overall Health
      1.    Visits per unique visitor                                        8
      2.    Page views per visit                                            10

      Driving Traffic to Multiple Audiences
      3.    Top entry and landing pages                                     12
      4.    Reject rate/bounce rate                                         13
      5.    Conversion rate for e-mail newsletters                          14

      Growth
      6.   Visitor frequency                                                15
      7.   Visitor recency                                                  16
      8.   New vs. returning visitors                                       17
      9.   Most popular stories                                             18

      Internal Search
      10.    Visits using internal search                                   19
      11.    Site exits after using internal search                         20

      Time Spent
      12.   Time spent during visits                                        21

Sources                                                                     22




_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                               1
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Using Web Analytics

Web analytics, or eMetrics, is the art and science of using research to analyze who visits a
website. There are two main types of research – behavioral and attitudinal.

Behavioral research measures what people actually did when they visited a site. “Actual
behavior” is one computer’s actions (the person’s computer) being captured by another (your
server).

Behavioral metrics measure:
      --      Did a person visit a page?
      --      If so, how many times did he or she visit?
      --      When did he or she visit?

Attitudinal research measures what people reported they did, and why. Usually done through
surveys, attitudinal research measures:
       --       what people are interested in, and whether they found it on your site
       --       where else they’re getting their news and information
       --       how satisfied they are with your site, and how likely they are to return


This document covers what newsrooms should use as a foundation for analyzing their
websites. Future versions of this document will include sections on Web 2.0 (e.g., video, user-
generated content), e-mail newsletters and engagement. Measuring engagement involves using
a combination of behavioral and attitudinal research metrics across both print and online
products. Fully understanding the scope and depth of commitment and emotional involvement
people have with your media organization is key to attracting, retaining and building current and
new audiences.




_____________________________________________________________________________
                         Dana Chinn, USC Annenberg School for Communication
                                                 2
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Developing a Web Analytics Plan

1.    Develop a strategic plan. Establish goals.
       Goals should be for specific audiences or content sections, not website elements like
         video or user-generated content.
       Define the actions that you think will lead to each goal.
       Ideally, your plan should be for both print and online.
         o    Editors are responsible for both, and making a decision for one often affects the
              other.
         o    Similarly, print and online research should be designed concurrently so you can
              understand reader behavior and attitudes across all experiences.

2.    Define the Key Performance Indicators for each goal. KPIs are the metrics by which
      you define progress.
       Define the actions you will take if a KPI goes up, and which to take if it goes down.

3.    Benchmark the KPIs. In other words, decide on a starting point so you’ll know how to
      interpret increases and decreases.
       Ensure your software can measure the KPIs, and that your site is coded correctly.
       Set a goal for each KPI, and establish a time period by which you expect to reach it.
          o Establish short-term (e.g., weekly, monthly) goals in addition to goals for each
              key point in the plan (e.g., after an event).

4.    Implement.

5.    Monitor each KPI. Address problems, and leverage successes.




_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                               3
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Defining Newsroom Key Performance Indicators

1.    Use KPIs for strategic planning, not for day-to-day decisions about coverage.
      However, track major news events or other factors that will skew a KPI temporarily but
      not reflect a change in the long run.

2.    A KPI is only useful if one person is held accountable for it.
      A significant increase or decrease in a KPI should immediately lead to an action led by
      one person.

      For a newsroom, this means:
            An editor should assign one person (e.g., assistant managing editor, section
              editor or senior reporter) to each section of a site.
            The person assigned to a section and its KPIs must have the authority, resources
              and tools to make and implement changes.

      This also means an editor of a newspaper site should only be held accountable for the
      KPIs for the sections he/she controls. Overall site KPIs for an editor should be based on
      a roll-up of the news/editorial sections; editors should not be evaluated based on overall
      site traffic numbers that include advertising-only content.

3.    KPIs are only useful if they’re actually used.
      Not all of the KPIs apply to each section or initiative. Select only those KPIs where it’s
      clear what question will be answered (and what action should be taken) if the KPI
      increases or decreases. Conversely, stop tracking a KPI if it’s no longer useful.

      Set the time periods based on how you’re going really use the information. Newspaper
      sites have new content daily, so you should review KPIs on a weekly basis at the
      minimum. However, a thorough monthly review that leads to decision-making is better
      than a hurried, cursory weekly review that leads to nothing.

4.    It’s better to base a decision on good data than to guess. However, it’s better to
      guess than to use bad data.
      Know the limitations of both behavioral and attitudinal research. If you have good data,
      you can be confident in your decision-making. If you have bad data, you might proceed
      the wrong way and commit more resources than you should. If you guess instead, you’ll
      proceed more cautiously and have more evaluation points.




_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                               4
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________


BASIC BEHAVIORAL METRICS
Below are the basic behavioral metrics used in the KPIs outlined in this document. By
themselves as simple counts (e.g., number of page views per month), they are not usually KPIs.
Counts don’t put numbers into context, so they don’t allow you to understand the dynamics of an
action enough to act. KPIs are usually percents or ratios.

Metric                         Definition                             Caveats

1.       Page view             Page on a website

2.       Unique visitor        A single visitor who may have          Measured with cookies, so
                               multiple visits in a defined           it’s really an inference.
                               reporting period, e.g.,                Examples: Will not capture a
                               day, week, month.                      person who uses multiple
                                                                      computers, or multiple people
                                                                      who use a single computer.

                                                                      Is always based on a time
                                                                      period, so can’t be added,
                                                                      e.g., each week in January
                                                                      can’t be added to get the
                                                                      monthly January number.

         New visitor           A unique visitor who has
                               visited a site for the first time

         Return visitor        A unique visitor who has visited
                               a site during a previous reporting
                               period

3.       Visit                 A single instance of any visitor       Based on a defined time
                               arriving at a website                  limit (usually 30 minutes).

4.       Entry page            The first page of a visit

5.       Landing page          The first page of a visit based        Usually used with a specific,
                               on a refer from a print source,        unique URL to track a
                               an ad, or a link within a site         campaign, e.g., more refers in
                                                                      Business print section; house
                                                                      ads for high school sports
                                                                      section.

6.       Bounce                A visit that has one page view

_____________________________________________________________________________
                          Dana Chinn, USC Annenberg School for Communication
                                                  5
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

ATTITUDINAL RESEARCH AREAS

Area                               Sample Survey Questions

1.     Satisfaction                “How often do you [read/contribute/buy] [content area]?”
                                   “How satisfied are you with [action/content area]?

2.     Loyalty                     “Please rate the level of loyalty you feel toward [website].”
                                   “How willing are you to recommend [content area]?”

3.     Brand equity                “Please indicate your level of agreement with the
                                   following statements. The [brand name]:
                                   --     Provides information on [content area] I can’t
                                          easily get elsewhere.
                                   --     Gives me the most complete understanding of
                                          [content area].
                                   --     Is more credible and trustworthy than most other
                                          news sources in [content area].”

                                   “What is the one specific source for [content area] you
                                   use the most?”

                                   “How often do you go to [brand name] for news and
                                   information about [content area]?

                                   “Please tell me whether you think the [brand name] is
                                   improving or declining in [content area].”
                                   “What things have you noticed that lead you to say the
                                   [brand name] is [improving/declining]?”

                                   “How often do you go to the [website] based on what
                                   you’ve seen in the [print brand name]?


4.     Print and online usage      Attitudinal research can capture how audiences feel about
                                   both your print and online editions. Although it doesn’t
                                   capture how people actually use both print and online, it
                                   does capture what people say they do. For strategic
                                   planning, an audience’s perception of how it uses content
                                   across all platforms is as equally important as actual
                                   behavior.




_____________________________________________________________________________
                      Dana Chinn, USC Annenberg School for Communication
                                              6
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Attitudinal research surveys using traditional statistical samples of the targeted audiences will
yield the most reliable and in-depth insights for strategic planning.

Be cautious in using research that doesn’t use statistical samples. For example, online pop-up
surveys and surveys based on a panel (a group of people interviewed repeatedly and periodically)
only give you data from people who volunteer information. Pop-up surveys also usually have
low overall response rates.

You can use information from pop-up and panel surveys to point out areas that need further
exploration. However, you won’t be getting a full and accurate picture of all of your current –
and potential – audiences, so you shouldn’t use them as the sole source of information for major
decisions.




_____________________________________________________________________________
                         Dana Chinn, USC Annenberg School for Communication
                                                 7
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Overall Health

1.       VISITS PER UNIQUE VISITOR

Formula:                               Number of visits for a specified time period
                                       Number of unique visitors for that time period

                                       Example: Number of visits the week of July 13
                                                Number of unique visitors the week of July 13


Caveats:                               Can’t roll up days into a week or weeks into a month. For
                                       example, you can’t take the weekly numbers of the five
                                       weeks in July to get a monthly July number.

                                       This is an average; you may want to use the median
                                       instead.


An INCREASE usually means:             Users are coming more frequently. Use frequency and
                                       recency KPIs (see the Growth KPIs) to get more detail.

                                       Major news events will skew this number and may lead to a
                                       false sense of improvement. Do attitudinal research to find
                                       out what caused this number to increase (so you can try
                                       more of the same).


A DECREASE usually means:              Users coming to a site less and becoming less engaged.
                                       There may be problems with
                                           the amount, quality or uniqueness of the content
                                           design and navigation
                                           refers from the print edition
                                           marketing efforts

                                       There may be new competitors, or content seasonality.


Newsroom questions:
   Is it realistic to expect a visit a day from users in your geographic market? Isn’t once a
      month or once a week too low?

        Does this metric increase significantly when you refresh content multiple times per day?


_____________________________________________________________________________
                          Dana Chinn, USC Annenberg School for Communication
                                                  8
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Visits Per Unique Visitor - Ratio Examples

                                     Number of
                                                                Visits per     Usually Good,
                                               Unique             Unique       Usually Bad,
                                        Visits Visitors            Visitor      or Depends
   Starting point or benchmark             30        10                  3

                  Unique
  Visits          Visitors
1 Increase        Increase                 150          30              5    Good
                                            50          20              3    Depends - watch
                                            60          60              1    Bad

                                                                             Bad - decrease in
2 Increase        Decrease                  40           5              8    visitors
3 Increase        Stay the same             60          10              6    Depends

4 Decrease        Increase                  20          20              1    Bad
5 Decrease        Decrease                  20           5              4    Bad
6 Decrease        Stay the same             20          10              2    Bad

  Stay the
7 same            Increase                  30          20              2    Bad
  Stay the                                                                   Bad - decrease in
8 same            Decrease                  30            5             6    visitors




_____________________________________________________________________________
                   Dana Chinn, USC Annenberg School for Communication
                                           9
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Overall Health

2.       PAGE VIEWS PER VISIT

Formula:                              Number of page views
                                        Number of visits


Caveats:                              This is an average, so it may be misleading.
                                      Use the median page views per visit if the content (or
                                      attitudinal research) suggests the number of page views per
                                      visit varies widely by user.


An INCREASE usually means:            Users are reading more content and becoming more
                                      engaged.

                                      Or, it may mean users are combing through your site and
                                      are frustrated that it takes too many clicks to find what
                                      they’re looking for.


A DECREASE usually means:             Users are looking at less content than before and becoming
                                      less engaged. There may be problems with
                                           the amount, quality or uniqueness of the content
                                           design and navigation
                                           refers from the print edition
                                           marketing efforts

                                      Or, you might have improved the navigation and/or internal
                                      search function.


Newsroom questions
   Print circulation and readership varies from day to day. But it’s easier and takes less time
      to access a website. Daily news consumption is different now. Thus, shouldn’t this KPI
      be relatively consistent regardless of the day of the week? That is, why should Monday’s
      page views per visit be significantly lower than Sunday’s?

        To attract a wide variety of audiences and to give them what they want and when they
         want it, should the amount of content put on the site be consistent from day to day?




_____________________________________________________________________________
                         Dana Chinn, USC Annenberg School for Communication
                                                10
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Page Views Per Visit - Ratio Examples

                                    Number of
                                                               Page           Usually Good,
                                       Page                views per          Usually Bad,
                                      views Visits              visit          or Depends
   Starting point or benchmark           30    10                  3

  Page views      Visits
1 Increase        Increase               150        30                  5   Good
                                          50        20                  3   Depends
                                                                            Bad - look at
                                           60       60                  1   bounce rate

                                                                            Bad - decrease in
2 Increase        Decrease                 40        5                  8   visits
                  Stay the
3 Increase        same                     60       10                  6   Depends

4 Decrease        Increase                 20       20                  1   Depends
                                                                            Bad - look at
5 Decrease        Decrease                 20        5                  4   bounce rate
                  Stay the
6 Decrease        same                     20       10                  2   Depends

  Stay the
7 same            Increase                 30       20                  2   Depends
  Stay the                                                                  Bad - decrease in
8 same            Decrease                 30        5                  6   visits




_____________________________________________________________________________
                   Dana Chinn, USC Annenberg School for Communication
                                          11
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Driving Traffic to Multiple Audiences

3.     TOP ENTRY AND LANDING PAGES

Formula:                            Number of times users entered your site through the home
                                    page
                                    Number of times users entered your site through other
                                    pages

                                    Percent of home page entry pages
                                    Percent of other page entry pages


Caveats:                            Need to have specific, easy-to-use URLs for as many
                                    sections as possible. Specific URLs are essential to
                                    tracking the effects of print refers, house ads and marketing
                                    campaigns.


Increases/decreases:                Decreases in the number or percent of home page entries
                                    should be accompanied by increases in the number or
                                    percent of section entry pages. This will indicate users are
                                    finding specific reasons to go to your website, which then
                                    may lead to a higher level of engagement.

                                    Home pages are essential orientation points for users.
                                    However, newspaper home pages are, by necessity,
                                    crowded with multiple navigation elements. This increases
                                    the possibility that users won’t find what they’re looking
                                    for. (See the Bounce Rate KPI on the next page.) “Too
                                    many” or “too high” of a percentage of home page entries
                                    may mean you should market (via print refers or marketing
                                    campaigns) a wider variety of content to a wider variety of
                                    audiences.


Newsroom questions
   What is the traffic increases that can be expected due to print refers?
   Do newspaper websites with the majority of users coming in through their home pages
      have lower levels of satisfaction and engagement than those with users coming in through
      the side doors?




_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                              12
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Driving Traffic to Multiple Audiences

4.    BOUNCE RATE

Formula:                           Number of bounces to designated page (e.g., home page)
                                    Number of visits beginning with the designated page


Caveats:                           Need to look at all high volume entry pages to fully
                                   understand increases and decreases to the home page.


An INCREASE usually means:         Users are attracted to your site or section, but aren’t finding
                                   what they’re looking for. Examine design, navigation and
                                   page loading times. Attitudinal research would probe for
                                   deeper reasons on where users who are interested in a
                                   particular topic are going, and why.


A DECREASE usually means:          You’ve made changes, and they’re working.

                                   Decreases in the home page bounce rate should be
                                   accompanied by increases in the number and percent of
                                   section entry pages.


Newsroom questions
   How does the use of a newspaper website home page vary by audience?
   Should only section and microsite URLs be promoted?




_____________________________________________________________________________
                      Dana Chinn, USC Annenberg School for Communication
                                             13
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Driving Traffic to Multiple Audiences

5.     CONVERSION RATE
        E-mail newsletter subscriptions
        RSS feeds
        Bookmarks

Formula:                            Number of unique visitors who subscribe
                                         Number of unique visitors


Caveats:                            Must be used with a specific time period. Thus, you can’t
                                    roll up days into a week or weeks into a month. For
                                    example, you can’t take the weekly numbers of the five
                                    weeks in July to get a monthly July number.

                                    Should be used together with cancelation rates.


An INCREASE may mean:               More users have found another way to engage with your
                                    content.


A DECREASE may mean:                There is a decrease in new visitors (see the New Visitors
                                    KPI in the Growth section).

                                    There may be problems with the subscription form itself.


Newsroom questions
   Are e-mail newsletters the most important tool newspapers have in driving traffic? For
      which audiences are they the most effective in increasing satisfaction and engagement?
   Will RSS feeds grow in importance? For what type of content?
   Can the use of bookmarks and the actual readership of RSS feeds (i.e., how often do
      people use a bookmark or read the RSS feed they’ve downloaded) be effectively
      measured with attitudinal or other research?




_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                              14
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Growth

6.    VISITOR FREQUENCY

Formula:                           Number of high/medium/low frequency unique visitors
                                               No. of unique visitors


Caveats:                           Need to define what’s high, medium and low by time
                                   period. Example:
                                          High frequency: 20+ visits per month
                                          Medium: 10-19
                                          Low: 1-10

                                   Need to have a granular list of each visit by visitor.


An INCREASE may mean:              An increase in the percent of high-frequency visitors
                                   indicates a website is probably building engagement and
                                   satisfaction.

                                   Attitudinal research would probe whether a site is building
                                   engagement but not satisfaction (sometimes seen in seven-
                                   day print subscribers).


A DECREASE may mean:               Decreases in low-frequency visitors should be accompanied
                                   with increases in medium- and high-frequency visitors.


Newsroom questions
   How can low- and medium- frequency visitors be encouraged to visit more? Is frequency
      affected by the amount of local and/or unique content?
   Which audiences are “naturally” low frequency? Is all traffic good, or are there low
      frequency audiences not worth pursuing?
   Attitudinal research can indicate whether there’s a gap between the percent of people
      who are news junkies vs. those who access your website daily.




_____________________________________________________________________________
                      Dana Chinn, USC Annenberg School for Communication
                                             15
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Growth

7.     VISITOR RECENCY

Formula:                            Number of high/medium/low recency unique visitors
                                                Number of unique visitors


Caveats:                            Need to define what’s high, medium and low within a
                                    specific time period. Example:
                                            High recency: Visited within the last three days. If
                                                   today is July 1, the visitor came to the site at
                                                   least once June 27-30.
                                            Medium: The visitor came to the site at least once
                                                   days 20-26 (e.g., June 20-26).
                                            Low: Last visited during days 1-20.

                                    Need to have a granular list of each visit by visitor.


An INCREASE may mean:               How recent a person has visited a site is a good predictor of
                                    whether he/she will return in the long run. Attitudinal
                                    research will also show whether high recency leads to
                                    increased satisfaction and engagement.

                                    Increases in low recency may mean there was a major news
                                    event earlier in the time period.


A DECREASE may mean:                Decreases in high recency visitors indicate people aren’t
                                    using your site as often as they used to. Attitudinal
                                    research is essential to determine whether this is due to
                                    problems with the site (e.g., content, design, navigation) or
                                    to unrelated reasons (e.g., major news events in the
                                    preceding time period; school no longer in session).


Newsroom questions
    How do recency predictors and definitions for high/medium/low differ:
         o For a newspaper with unique content vs. a news site with wire (e.g., Yahoo!)?
         o By content?
         o Audience and geography?
    How does print readership affect website recency?


_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                              16
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Growth

8.     NEW VS. RETURNING VISITORS
        Percent of new visitors
        Percent of returning visitors

Formula:                            Number of new unique visitors in a time period
                                      Number of unique visitors in a time period

                                    +

                                    No. of returning unique visitors in a time period
                                       No. of returning visitors in a time period

                                    = 100%


Caveats:                            Can’t roll up weeks into a month or months into a year.
                                    For example, you can’t take the weekly numbers of the five
                                    weeks in July to get a monthly July number.


Increases/Decreases:                The goals for the percent of new or returning visitors
                                    should be set based on the goals of specific initiatives.

                                    An increase in the percent of new visitors will result in a
                                    decrease in returning visitors, and vice versa.

                                    Example:
                                    When a new section for high school sports is launched, the
                                    percent of new visitors should increase, and the percent of
                                    returning visitors should decrease. (Attitudinal research
                                    can indicate whether new visitors were in the targeted
                                    audience of high school students.)


Newsroom questions
    What is the percent of new visitors a newspaper website should have over the next two
       years to build a solid foundation?
    How do the expectations for the percent of new visitors differ by geographic market and
       audience?
    How do these percents vary based on how much content is in the print edition?



_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                              17
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Growth

10.    MOST POPULAR STORIES

Formula:                             Develop a scoring system for measuring stories that:
                                         are the most-read (page views)
                                         have attracted the most comments
                                         readers have recommended or e-mailed the most


Caveats:                             This KPI can be valuable, but is extremely labor intensive
                                     and subject to misinterpretation. It should mainly be used
                                     to in conjunction with other KPIs to give additional insight.

                                     “Most read” is difficult to define with page views. Much
                                     depends on your site design. For example, do you count
                                     whether all pages of a story were viewed, or just the first
                                     page, or somewhere in between?

                                     When measuring the number of comments, you develop
                                     rules on what type of comments should be eliminated from
                                     the count.


Increases/Decreases:                 This KPI is probably best used when you’ve launched a
                                     major new or improved section, and have implemented a
                                     corresponding, comprehensive marketing campaign (e.g.,
                                     print and online refers, multimedia marketing).


Newsroom questions (see “The Pleasure Principle,” by Michael Hirschorn, Atlantic, Dec. 2007)
   In the short-term, stories on popular national subjects like Britney Spears generate traffic
      spikes. However, in the long run, do these types of stories build lasting audiences for
      community newspaper websites? Will local unique stories eventually be the most
      popular?
   Are these counts consistent enough to be used to predict reader behavior?
   How much do what readers think is interesting overlap with what editors think is
      important?
   What type of stories get low traffic but are essential to a news organization’s credibility
      and thus, contribute indirectly to traffic? This is the Netflix business model.
      Subscribers come to Netflix due to the depth of their movie inventory. However, most
      subscribers only rent the blockblusters.



_____________________________________________________________________________
                        Dana Chinn, USC Annenberg School for Communication
                                               18
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Internal Search

10.    VISITS USING INTERNAL SEARCH
        Visits using internal search
        Search terms used in internal search; no. of “no results” from common terms


Formula:                            Number of visits that result in a search results page
                                                   Number of visits


Caveats:                            These metrics are useful if several elements are in place,
                                    including:
                                         good understanding of current and future audience
                                           content interests and needs
                                         content classification process and system
                                         an understanding of how paid archived content
                                           shows up in internal searches

                                    These metrics are a starting point of a more extensive effort
                                    toward optimizing both internal and external search (e.g.,
                                    optimizing your site for Google).


An INCREASE may mean:               The internal search box is new, or was moved to a more
                                    visible or usable place.

                                    Or, an increase in the number of searches for common
                                    search terms (e.g., “movies”) may mean users are having
                                    problems with basic navigation (e.g., using the print section
                                    name of “Preview” for movies).


A DECREASE may mean:                Basic navigation was improved.

                                    Or, the search box might have been buried. Also, a section
                                    may lack depth in its content, so searches aren’t needed.


Newsroom questions
   How important is internal search in increasing engagement and satisfaction?
   Should editors spend more time and resources on design, navigation and usability than on
      internal search?


_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                              19
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Internal Search

11.    SITE EXITS AFTER USING INTERNAL SEARCH


Formula:                             Number of site exits from search results pages
                                       Number of visits to search results pages


Caveats:                             Should be used together with Visits From Internal Search.
                                     Can’t be used if a site’s internal search engine leads users
                                     off the site.


An INCREASE may mean:                Your search engine isn’t returning the desired results, and
                                     visitors are leaving the site because they can’t find what
                                     they’re looking for – and what they expected to find on
                                     your site with common search terms.

                                     Search engines are gateways to additional content and
                                     increasing visitor satisfaction. Attitudinal research can
                                     probe more into what visitors were expecting, and with
                                     which search terms.


A DECREASE may mean:                 Your search engine has improved in its ability to keep
                                     visitors on your site.


Newsroom question
   Should a newspaper website lead visitors off its site through its internal search engine, if
      visitors expect from a news organization to be comprehensive?




_____________________________________________________________________________
                        Dana Chinn, USC Annenberg School for Communication
                                               20
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________

Key Performance Indicators: Time Spent

12.   TIME SPENT DURING VISITS

Formula:                            Number of high/medium/low time-spent visits
                                                Number of visits


Caveats:                            Need to define what’s a high, medium or low time spent on
                                    a visit. Example:
                                             High-time spent: more than five minutes in a visit
                                             Medium: between 30 seconds and five minutes
                                             Low: less than 30 seconds

                                    Need to have a list of each visit by each visitor.

                                    Metrics using time spent don’t answer many questions (and
                                    thus, don’t directly lead to decisions) because time spent
                                    varies too much based on content and the user. “Average
                                    time spent per visit,” a common measure seen in many
                                    reports, is a particularly unhelpful number due to these
                                    extreme variations.

                                    Frequency and recency metrics are better behavioral
                                    metrics for inferring engagement and predicting future
                                    behavior. If used, time-spent metrics should be used only
                                    as a supplement to other metrics. Attitudinal research can
                                    show whether time reported by users mirrors actual
                                    behavior, and, most importantly, whether users are
                                    spending a lot of time on the site because they want to.


Increases/Decreases:                An increase in the percent of visits of five minutes or more
                                    may mean users are engaged in your content. Or, it may
                                    mean they can’t find what they want.

                                    An increase in the percent of visits of less than 30 seconds
                                    may mean users didn’t find what they want. Or it may
                                    mean they immediately found it.


Newsroom questions
   How does time spent vary by audience? By type of content?
   What are the time-spent thresholds for building engagement?

_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                              21
Measuring Web Success in the Newsroom
Web Analytics for Newspaper Websites – Version 1 (July 2008)
______________________________________________________________________________


Sources

Dana Chinn
Lecturer, School of Journalism
USC Annenberg School for Communication
E-mail: chinn@usc.edu
Phone: 213-821-6259


Web Analytics Overviews
   “Web Analytics: An Hour a Day,” by Avinash Kaushik, Wiley Publishing, 2007

      “Actionable Web Analytics: Using Data to Make Smart Business Decisions,” by Jason
       Burby & Shane Atchison, Wiley Publishing, 2007

      ClickZ Marketing Analytics e-mail newsletters, various columnists
       www.clickz.com


Metric Details and Definitions
   “Web Analytics Demystified,” by Eric Peterson, Celilo Group Media and CafePress,
      2004

       “The Big Book of Key Performance Indicators,” by Eric Peterson
             Available at www.webanalyticsdemystified.com

      “Web Analytics Definitions – Version 4.0,” Web Analytics Association, August 16, 2007
       www.waa.org


Conference
    eMetrics Marketing Optimization Summits – conferences and boot camp sessions co-
      sponsored by the Web Analytics Association. These conferences, which previously
      focused on e-commerce, now include a “content-driven” track. The 2008 San Francisco
      eMetrics featured a presentation from the New York Times.
      www.emetrics.org




_____________________________________________________________________________
                       Dana Chinn, USC Annenberg School for Communication
                                              22

				
DOCUMENT INFO